Spatial Relationship between Cerebral Microbleeds, Moyamoya Vessels, and Hematoma in Moyamoya Disease Ken Kazumata, MD,* Daisuke Shinbo, MD,* Masaki Ito, MD,* Hideo Shichinohe, MD,* Satoshi Kuroda, MD,† Naoki Nakayama, MD,* and Kiyohiro Houkin, MD*
Background: Adult moyamoya disease (MMD) is known to have high incidence of cerebral microbleeds (cMBs); however, the clinical significance still remains unclear. We investigated the frequency of cMBs in a large number of patients and analyzed the patterns of MB distribution in association with the location of the hematoma and moyamoya vessels. Methods: We studied 259 consecutive patients with MMD using prospectively collected database. One hundred ninety-one patients were eligible for the present study, and image analysis was performed retrospectively. The presence of cMBs and remains of hemorrhage were determined using gradient-echo T2*weighted sequence (1.5 T). The development of moyamoya vessels was assessed on source images of time-of-flight magnetic resonance angiography. The analysis consists of descriptive assessment of the spatial relationship between cMB, remains of hemorrhage, and moyamoya vessels. Statistical analysis was performed to calculate relative risk ratio in the presence of cMBs in relation to the remains of hemorrhage (macrohematoma), age of onset, and the presence of concomitant moyamoya vessels. Results: Thirty MBs were observed in 20 adult MMD patients (16.9%). MBs were located predominantly in the periventricular white matter (63.3%) followed by the basal ganglia/thalami (20%). Comparing the patients with cMBs from those without, hematoma was more frequently observed in patients with cMBs (odds ratio [OR] 4.29; 95% confidence interval [CI] 1.58-11.62; P 5 .0062). Patients with adult onset was more likely to demonstrate cMBs (14.4%) compared with the patients with pediatric onset (4.1%) (OR 3.93; 95% CI 1.11-13.91). Moyamoya vessels appeared in the lateral part of the trigon, and the periventricular white matter was significantly associated with the presence of cMBs (lateral part of the trigon; OR 3.29 [1.59-6.82], P 5 .0019, periventricle of the body of lateral ventricle; OR 2.40 [1.20-4.79], P 5.0214, respectively). cMBs accompanied concomitant arteries in 23 (76.7%) lesions. The subependymal–leptomeningeal artery anastomosis was the most common pattern (n 5 20, 66.7%). Conclusions: Spatial relationship was demonstrated between the moyamoya vessels and perivascular hemosiderin deposition particularly around the subependymal–leptomeningeal anastomosis, suggesting the mechanism for the development of cMBs in MMD. Present study further supports previous findings that cMBs potentially serve as a marker for the bleeding-prone microangiopathy in MMD. The significance of the present study lies in selecting optimal surgical candidate for preventing future hemorrhage by the presence of the cMBs, whereas current surgical indication relying on the degree
From the *Department of Neurosurgery, Hokkaido University Graduate School of Medicine, Sapporo; and †Department of Neurosurgery, Toyama University Graduate School of Medicine, Toyama, Japan. Received July 22, 2013; revision received October 10, 2013; accepted December 3, 2013. Funding: No grants or funding sources are pertinent to this article. Conflicts of interest: None.
Address correspondence to Ken Kazumata, MD, Department of Neurosurgery, Hokkaido University Graduate School of Medicine, North 15 West 7, Kita, Sapporo 060-8638, Japan. E-mail: kazumata@ med.hokudai.ac.jp. 1052-3057/$ - see front matter Ó 2013 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2013.12.007
Journal of Stroke and Cerebrovascular Diseases, Vol. -, No. - (---), 2013: pp 1-8
1
K. KAZUMATA ET AL.
2
of ischemia frequently fails to detect patients with future hemorrhage. Key Words: Moyamoya disease—microbleeds—intracranial hemorrhage—intraventricular hemorrhage—subependymal artery—medullary artery. Ó 2013 by National Stroke Association
Introduction MMD is a progressive steno-occlusive cerebrovascular disease that is characterized by the presence of net-like collateral vessels (moyamoya vessels) at the base of the brain.1 Recent reports demonstrated that the incidence of cerebral microbleeds (cMBs) was higher in patients with MMD compared with healthy subjects.2-6 In addition, the presence of cMBs predicts the future risk of hemorrhage.2,7 These results suggest that cMBs can serve as a marker of bleeding-prone microangiopathy in MMD. A certain pattern of distribution of MBs serves to indicate microangiopathy developed by the different mechanisms. cMBs in the basal ganglia and cortical– subcortical distribution are characteristic in hypertensive lipohyalinosis and amyloid angiopathy, respectively. Similarly, it is speculated that MMD may demonstrate specific pattern of the distribution of cMBs and follow the spatial distribution of hematoma and moyamoya vessels. The investigations of these spatial relationships may further elucidate the clinical significance and the mechanism for the development of the cMBs in MMD. In addition to the spatial distribution of cMBs, the exact number or locations of the bleeding have also not been fully evaluated previously. T2* gradient-echo or susceptibility weighted image not only demonstrate cMBs but also reveals remains of tiny intracerebral hemorrhage throughout life, which enables determining the cause of the hemorrhage years after the bleeding event. This is particularly useful to determine the location of the hemorrhage in primary intraventricular hemorrhage, a frequent manifestation of the intracranial bleeding in adult MMD. Furthermore, the source image of magnetic resonance angiography (MRA) demonstrates the enlarged basal perforators and choroidal arteries developed as a collateral circulation, which potentially reveal spatial association between the moyamoya vessels and the cMBs. Present study attempted to provide additional information by performing an analysis of the frequency and distribution patterns of MBs in association with the hematoma location and moyamoya vessels in a large consecutive series of patients with MMD.
Patients and Methods We examined 259 consecutive patients with MMD at the Department of Neurosurgery of the Hokkaido University Hospital from March 1980 to March 2012. The diagnosis was based on the guidelines established by the Research Committee on Moyamoya Disease (Spontaneous Occlu-
sion of the Circle of Willis) of the Ministry of Health and Welfare of Japan.8 These patients were referred to our hospital for the diagnosis and the treatment. Demographic information, including history of stroke, medical history, and medications for hypertension, diabetes mellitus, or hyperlipidemia, was collected. We also recorded the prevalence of antithrombotic therapy before the occurrence of recurrent stroke in each patient. Their information is tabulated in the database prospectively and used in this study to select eligible patients in this study. Since 2003, a standardized magnetic resonance imaging (MRI) protocol has been employed for all the time points that include fluid-attenuated inversion recovery (FLAIR), T2, and gradient recalled echo (GRE) sequences and MRA.5 Gradient-echo T2*-weighted sequence was available for assessment in 191 patients (73.7%; 56 men and 135 women; age, 2-86 years; mean age, 40.0 1 19.3 years). One hundred forty-one of the 191 patients underwent revascularization surgery (73.8%). In turn, 70 of these 141 patients underwent gradient-echo T2*-weighted sequence studies before revascularization surgery, and the remaining 71 underwent imaging after revascularization surgery. No patient had a known pathogenesis, such as head trauma, cerebral amyloid angiopathy, current anticoagulant therapy, arteriovenous malformation, cavernoma, or other systemic coagulation disorder.
Imaging Analysis MRI was performed every 1 year and/or at the time of recurrent stroke using a 1.5 T MR scanner (Magnetom Vision; Siemens Medical Solutions, Erlangen, Germany). The imaging protocol consisted of the axial spin-echo T1-weighted imaging (WI) (repetition time/echo time [TR/TE] 5 600/14 ms, number of excitation [NEX] 5 1, matrix size 5 512 3 512), axial fast spin-echo T2WI (TR/TE 5 4000/96 ms, effective echo train length 5 7, NEX 5 1, matrix size 5 512 3 512), and axial gradientrecalled echo T2*WI (TR/TE 5 800/26 ms, flip angle 5 20 , NEX 5 1, matrix size 5 256 3 256). All images were acquired with a 240-mm field of view, a 5.0mm section thickness, and a 1.5-mm intersection gap. Axial source images acquired by the 3-dimensional time-of-flight magnetic resonance angiography (3DTOF-MRA) (TR/TE 5 27/7.2 ms; flip angle 5 20 ; NEX 5 1, field of view 5 230 mm; recon matrix 5 512 3 512; number of slabs 5 4 [180 sections]; section thickness 5 .65 mm; acquisition time 5 9 min) were used for identifying moyamoya vessels.
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
Gradient-echo T2*-weighted sequence is also sensitive in detecting small remains of previous parenchymal hemorrhage in primary intraventricular hemorrhage. Definite cMBs were defined as rounded foci of hypointense signal on GRE sequences with a diameter of 2-10 mm and not well seen on T2WI.9 Microbleed mimics were carefully excluded. Vessels with linear or curvilinear lesions in the subarachnoid space, with a cortical or juxtacortical location, were excluded. We did not include lesions that were adjacent to the infarction or hemorrhage or the area of surgery. Areas of symmetric hypointensity in the globus pallidus or dentate nuclei on GRE MRI were considered to likely represent calcifications and were not included. Air– bone interfaces in the frontal and temporal lobes were excluded from the analysis. The topographical distribution was demonstrated on the basis of the Microbleed Anatomical Rating Scale.9 The diagnosis of moyamoya disease (MMD) was achieved using MRA in recent patients. Therefore, arterial involvement was described by the composite score of the rating based on TOF MRA published elsewhere instead of Suzuki’s grading.10 Moyamoya vessels with linear or curvilinear structures in the basal ganglia, thalamus, and periventricular white matter were assessed on the source image of TOF MRA. The presence was described by either ‘‘visible’’ or ‘‘not visible.’’ No attempt of quantitative assessment was carried out in terms of vascular density of the moyamoya vessels. For the current analysis, 2 investigators performed the imaging evaluations using most recent MRI acquisition in each individual. Presence of cMBs and hematoma and the location were investigated by 1 researcher (D.S.), whereas rating the scale of MRA and the location of moyamoya vessels were investigated by the other researcher (K.K.). We further consulted previous MRI acquisition to determine whether these are the de novo lesions. The process of data collection was blinded between the 2 investigators. After completing data collection, we assessed the radiological rating, and any disagreement between these authors was resolved in consensus meeting with a third author (I.M.).
Statistical Analysis Differences in dichotomous variables were analyzed using Pearson chi-squared analysis or Fisher exact test. Student t test or the ANOVA was used to analyze differences in the mean or median of continuous variables between groups.
Results Demographic Data The population used here included 118 patients with adult MMD (age at onset, .18 years) and 73 patients with pediatric MMD (age at onset, #18 years). The cohort included presence of hematoma (n 5 31), presence of
3
cMBs alone (n 5 12), and absence of hematoma and cMBs (n 5 148). The average follow-up period was 104.0 6 92 months. Thirty cMBs were detected in 20 patients (age range 2365 years) on gradient-echo T2*-weighted sequence, corresponding to 16.9% in adult population; 15 cMBs were found in 8 patients with hematoma, whereas the remaining 15 cMBs were found in 12 patients without hematoma (ischemic symptoms [n 5 7], asymptomatic [n 5 4], and headache [n 5 1]; P , .01). Comparing the patients with cMBs from those without cMBs, hematoma was significantly more frequent in the patients with cMBs (odds ratio [OR] 4.29; 95% confidence interval [CI] 1.58-11.62; P 5 .0062). Patients with adult onset was more likely to demonstrate cMBs (14.4%) compared with the patients with pediatric onset (4.1%) (OR 3.93; 95% CI 1.11-13.91). The clinical parameters of patients in the groups with cMBs and no cMBs are listed in Table 1.
Topographic Patterns of cMBs and Hematoma In total, 41 macrohemorrhage (hematoma and tiny remains of hemorrhage at the paraventricle) lesions in 31 patients were identified (Fig 1, Table 2). Primary ventricular hemorrhage was documented in 8 patients before using gradient-echo T2*-weighted sequence. Small remains of hemosiderin on gradient-echo T2*-weighted sequence identified the origin of intraventricular hemorrhage (IVH) as being subependyum of the trigon (n 5 1), the lateral ventricle (n 5 2), and the temporal horn (n 5 1). cMBs were subependymal to the lateral ventricle in the lateral part of the trigon or the posterior two thirds of the periventricle (deep and periventricular white matter) in 19 lesions (63.3 %). Lobar microbleeds were observed in 3 lesions. Locations classified according to the Microbleed Anatomical Rating Scale are demonstrated in Table 2 and Figure 1. A trend for susceptibility to bleeding was observed in the basal ganglia and corpus callosum (P 5 .0184). The deep and periventricular white matter was susceptible to the development of cMBs (P 5 .07).
Moyamoya Vessels on Source MRA Images and Their Relationship with cMBs and Hemorrhage On MR, the linear structures were oriented in the known direction of perforating arteries (Figs 2-4). The source images of MRA were available in total of 298 hemispheres (cMBs [n 5 38], no cMBs [n 5 260]). In adult, moyamoya vessels were visible in the basal ganglia (n 5 78, 26.2%), thalamus (n 5 47, 15.8%), lateral part of the trigon (n 5 58, 19.5%), and periventricular white matter (n 5 89, 29.9%). Moyamoya vessels appeared in the lateral part of the trigon, and the periventricular white matter was significantly associated with the presence of cMBs (lateral part of the trigon; OR 3.29 [1.59-6.82], P 5 .0019, periventricle of the body of
K. KAZUMATA ET AL.
4
Table 1. Demographic and clinical variables for patients with cMBs and without cMBs Variables
cMBs (n 5 20)
No cMBs (n 5 171)
Age, y (mean 6 SD) Male, n (%) Symptom, n (%) Headache TIA Infarction Surgery, n (%) Onset age . 18 y, n (%) Medication Hypertension Hyperlipidemia Antiplatelet Hematoma, n (%) MRA composite score*
39.0 6 19.7 4 (20.0)
48.0 6 12.6 47 (27.5)
0 7 (35.0) 1 (5.0) 10 (50.0) 17 (85.0)
7 (4.1) 77 (45.0) 45 (26.3) 116 (67.8) 101 (59.1)
6 (30.0) 1 (5.0) 3 (15.0) 8 (40.0) 10.5 6 3.9
32 (18.7) 11 (6.4) 21 (12.3) 23 (13.5) 11.3 6 4.6
OR
P value ,.001 NS
3.93 (95% CI 1.11-13.91)
NS NS NS NS .0028
4.29 (95% CI 1.58-11.62)
NS NS NS .0062
Abbreviations: CI, confidence interval; cMBs, cerebral microbleeds; MRA, magnetic resonance angiography; NS, nonsignificant; OR, odds ratio; TIA, transient ischemic attack. *Arterial involvement rated by using MRA10
lateral ventricle; OR 2.40 [1.20-4.79], P 5 .0214, respectively). The typical patterns of anastomosis were identified in relationship with the moyamoya vessels as follows:
a) Distal part of the lateral posterior choroidal artery terminates in subependymal artery (SEA). SEA penetrated the lateral part of the trigon and the mid-posterior part of the lateral wall of the lateral
Figure 1. The patterns of distribution of microbleeds (open circles) and hemorrhage (closed circles) in moyamoya disease were depicted on the Microbleed Anatomical Rating Scale template.9
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
Table 2. The number of cMBs and the hemorrhage classified based on the MARS
Basal ganglia Thalamus Internal capsule External capsule Corpus callosum Deep and periventricular white matter Frontal Parietal Temporal Occipital Insula Brain stem Cerebellum
MBs
Hemorrhage
2 2 2 0 0 19
13* 3 4 0 7* 12y
2 1 0 0 2 0 0 30
0 0 1 0 0 1 0 41
Abbreviations: cMBs, cerebral microbleeds; MARS, Microbleeds Anatomical Rating Scale. Prevalence of cMBs and hematoma were compared across the regions. Hematoma is frequently observed in the basal ganglia and the corpus callosum, whereas cerebral microbleeds is relatively rare in these locations. cMBs are frequently observed in the deep and periventricular white matter (P 5.63), whereas hematoma in deep and periventricular white matter is less frequent manifestation (P 5.29). *P , .05, statistically significant. yP , .01, statistically significant, by the Fisher exact test.
ventricle. The linear signal took a radiating course toward the deep sulci (medullary artery), where leptomeningeal arteries in the temporoparietal region were identified (Fig 2).
Figure 2. Typical periventricular microbleeds and subependymal–leptomeningeal anastomosis. (A) Gradient-echo T2*-weighted imaging at the level of the lateral ventricles. (B and C) The source image of time-of-flight magnetic resonance angiography shows a linear high signal in the subependymal zone, deep white matter (short arrow), and deep sulci, which is consistent with a collateral pathway that developed because of anastomosis between subependymal artery and the leptomeningeal artery. (D and E) The coronal images in different patient demonstrate the course of the enlarged medullary arteries developed in bilateral parietal regions.
5
b) Thalamoperforating or medial posterior choroidal artery penetrated the lateral wall of the third ventricle. These anastomotic arteries traverse the basal ganglia and thalami, generating an anastomosis with either the long insular artery or the medullary artery (Fig 3). c) Anastomosis between the lenticulostriate artery and the medullary artery (Fig 3). In 23 (76.7%) cMBs, moyamoya vessels are identified lying next to them. Of these, anastomosis between SEA and the leptomeningeal artery was identified in 20 cases (66.7%). In 4 of the basal ganglia and thalamic cMBs (striatal; n 5 3, thalamus; n 5 20), a concomitant artery was identified in 4 cases (SEA–medullary artery, n 5 3; lenticulostriate artery, n 5 1). Source images of TOF MRA revealed concomitant arteries lying next to the remains of hematoma in 17 of 29 patients (58.6%) and concomitant arteries in 11 lesions of 9 patients with striatal hemorrhage. SEA–medullary anastomosis was identified in 5 of 8 patients with primary intraventricular hemorrhage. Source images of TOF MRA acquired before the hemorrhagic event were available for 8 of 10 cases of hemorrhage and 5 of 6 de novo cMBs. Retrospective image assessment identified moyamoya vessels at the site of 6 of 8 subsequent hemorrhage and 5 of 5 de novo cMBs.
Discussion In this study, we investigated the prevalence and distribution pattern of the cMBs among the largest number of patients with MMD investigated to date. The 17% rate
6
Figure 3. Patterns of moyamoya vessels. The source image of magnetic resonance angiography demonstrates the spatial relationship of the basal ganglia, lateral ventricle, and moyamoya vessels in coronal views. The lenticulostriate artery develops and forms an anastomosis between the medullary artery at the corner of the lateral ventricle on the right hemisphere (arrow; upper image). Bilateral thalamoperforating arteries give rise to the collateral network with long insular and medullary arteries on the left hemisphere (small arrow; middle 2 images). Choroidal arteries develop and form anastomosis with the medullary arteries at the lateral corner of the trigon (small arrow; bottom image).
of cMBs in MMD was lower than our preceded prospective cohort study in the nontreated MMD. We extended our survey to obtain greater population even in patients with a history of the revascularization surgery. The patients with cMBs were observed predominantly in adult-onset MMD. Posterior two thirds of the periventricular white matter of the lateral ventricle was the most common location of the cMBs. Presence of the periventricular moyamoya vessels and the hematoma is associated with the greater frequency of the cMBs. In addition, the location of the periventricular cMBs were considered to correspond with the site of anastomosis developed between the SEA and the medullary artery. cMBs are iron or hemosiderin deposits in the macrophages located in the perivascular space adjacent to small arteries, arterioles, and capillaries.11,12 It is considered as a marker of microangiopathy and speculated as a smallest form of the hemorrhage in some investigators.13-15 The average age of patients was 44 years old with cMBs,
K. KAZUMATA ET AL.
10 years younger than that of patients with the history of hemorrhage (54 years old). cMBs were more frequent in patients with hematoma than in patients without. Adult patients who underwent revascularization in childhood were less likely to demonstrate cMBs. These results suggest that cMBs may develop in adult MMD with certain disease duration and are associated with the later hemorrhage. The revascularization surgery might have prevented from developing these bleedingprone microangiopathy.16 The study of the distribution may further elucidate the clinical significance of cMBs. The periventricular predominant cMBs is considered characteristic distribution of MMD because deep periventricular cMBs are rare in stroke patients with other etiologies.9 In MMD, intraventricular hemorrhage is a frequent manifestation of the intracranial bleeding. Previously, the microaneurysm arising at the choroidal arteries have been considered responsible for such bleeding.17-19 Gradient-echo T2*weighted sequence revealed tiny remains of hematoma in periventricle in patients with a history of the primary hemorrhage. This suggests that the cause of IVH has 2 distinct mechanisms: one is the rupture of intraventricular aneurysms and the subependymal hemorrhage ruptured into the ventricle in the other. The latter is considered to take more benign prognosis than the former because the performance state in patients with IVH was not as poor as previously reported.20 The hemorrhage in the basal ganglia is observed in approximately half of the intracranial hemorrhage, whereas the cMBs in the basal ganglia and thalami were observed in only 20%. This discrepancy indicates that a certain population of hemorrhage in the basal ganglia develops spontaneously. In addition, it is speculated that the culprit vessels for the hemorrhage and cMBs may not be entirely the same in the basal ganglia. We further investigated the spatial relationship between the moyamoya vessels and the cMBs. The medullary artery irrigates white matter above the level of the lateral ventricle via the leptomeningeal artery.21 In particular, cortical arteries located around the watershed zone between the anterior cerebral and middle cerebral artery give rise to the medullary arteries.22 Lenticulostriate and long insular arteries irrigate white matter below the level of lateral ventricle.23,24 A collateral network via the deep perforating arteries or choroidal arteries develops when the cortical arteries were involved with disease progression. Lenticulostriate, thalamoperforating, and medial posterior choroidal arteries participate in the collateral networks with medullary arteries in the frontal region. Lateral posterior choroidal artery participates in the collateral network in the temporoparietal region. These anastomotic networks develop most exclusively at the lateral corner to the lateral ventricle. In patients with advanced age, temporoparietal region is susceptible for the ischemia.25 Therefore, a collateral network between
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
7
Figure 4. Illustrations of the hemorrhagic pattern attributed to bleeding from the subependymal artery (SEA)–leptomeningeal anastomosis. (A) Subependymal hemorrhage can be regarded as an enlarged form of periventricular cerebral microbleeds (cMBs). (B) A case of intraventricular hemorrhage. Remains of hemorrhage can be traced along the medullary artery. (C) cMBs in the posteroventral putamen developed along the course of SEA–medullary anastomosis. (D) Subcortical hemorrhage in the temporal lobe that developed because of bleeding from the SEA–medullary anastomosis.
the posterior choroidal and the medullary artery is considered developing predominantly in adult patients with advanced age. This may explain that cMBs develop most frequently at the site of anastomosis taking place between the choroidal arteries, deep perforators, and the medullary arteries. A 5- to 30-fold increase in diameter is estimated when SEA or the medullary artery are visible on MRI as the size of the choroidal arteries is 190 mm on average (range 40-490 mm) and the size of arterioles is 20-30 mm.26 Pathological studies of collateral circulation have demonstrated the presence of hemodynamic stress in MMD, revealing the breakdown of the elastic lamina and the thinning of the media.27 The pattern of collateralization via the SEA can also be described as a ventriculofugal and ventriculopetal anastomosis.28-30 MMD is the first of its kind to demonstrate such anastomosis clearly in living human imaging. The pattern of hematoma observed in MMD suggests that SEA–leptomeningeal anastomosis potentially induce a wide array of hemorrhagic lesions located in the subependymal zone, ventricle, posterior ventral putamen, and temporal subcortical region (Fig 4). The limitations of this study include its retrospective nature and the modality used, which is less sensitive compared with the currently available high-field MR scanner.4 In contrast with previous prospective studies conducted using a high-field MR scanner, this retrospective study revealed a smaller incidence of cMBs in
MMD.2,4,5,7 It should be noted that the signal intensity observed in blood vessels, as imaged by threedimensional TOF MRA, is dependent on the rate of blood flow. The frequency accompanying a concomitant artery adjacent to cMBs is, thus, potentially inaccurate and may be increased in studies using a high-field modality. It is necessary to differentiate gradient-echo T2*weighted hypointense lesions from cMB mimics, such as small vessels, calcification of small arteries, distended dissecting arteries, thrombosed arterioles, and capillaries. In particular, subependymal cMBs can be seen as an intraventricular lesion that mimics hemosiderin deposition that occurs in the choroid plexus after intraventricular hemorrhage. Therefore, some of these subependymal cMBs may be regarded as probable cMBs.9,31 However, considering the diameter of the cMBs, which was larger than the size of the accompanying vessels, and the absence of flow void on T2WI, gradient-echo T2*weighted hypointense lesions are consistent with the findings of cMBs.9,32
Conclusions The prevalence and distribution of cMBs were studied in a large number of patients with MMD. The distribution of cMBs in MMD showed a strong predilection for the subependymal zone of the posterior two thirds of the
K. KAZUMATA ET AL.
8
lateral ventricle. Collateral circulation between SEA and the medullary artery was demonstrated in close proximity to cMBs, suggesting that anastomosis between the ventriculopetal and ventriculofugal arteries is associated with the characteristic distribution of cMBs in MMD.
References 1. Suzuki J, Takaku A. Cerebrovascular ‘‘moyamoya’’ disease. Disease showing abnormal net-like vessels in base of brain. Arch Neurol 1969;20:288-299. 2. Kikuta K, Takagi Y, Nozaki K, et al. Asymptomatic microbleeds in moyamoya disease: T2*-weighted gradientecho magnetic resonance imaging study. J Neurosurgery 2005;102:470-475. 3. Kikuta K, Takagi Y, Nozaki K, et al. Histological analysis of microbleed after surgical resection in a patient with moyamoya disease. Neurol Med Chir 2007;47:564-567. 4. Kikuta K, Takagi Y, Nozaki K, et al. The presence of multiple microbleeds as a predictor of subsequent cerebral hemorrhage in patients with moyamoya disease. Neurosurgery 2008;62:104-111. discussion 111-102. 5. Ishikawa T, Kuroda S, Nakayama N, et al. Prevalence of asymptomatic microbleeds in patients with moyamoya disease. Neurol Med Chir 2005;45:495-500. discussion 500. 6. Kuroda S, Kashiwazaki D, Ishikawa T, et al. Incidence, locations, and longitudinal course of silent microbleeds in moyamoya disease: a prospective t2*-weighted MRI study. Stroke 2013;44:516-518. 7. Mori N, Miki Y, Kikuta K, et al. Microbleeds in moyamoya disease: susceptibility-weighted imaging versus T2*-weighted imaging at 3 Tesla. Invest Radiol 2008; 43:574-579. 8. Fukui M. Guidelines for the diagnosis and treatment of spontaneous occlusion of the Circle of Willis (’moyamoya’ disease). Research committee on spontaneous occlusion of the Circle of Willis (moyamoya disease) of the ministry of health and welfare, Japan. Clin Neurol Neurosurgery 1997;99(Suppl 2):S238-S240. 9. Gregoire SM, Chaudhary UJ, Brown MM, et al. The Microbleed Anatomical Rating Scale (MARS): reliability of a tool to map brain microbleeds. Neurology 2009; 73:1759-1766. 10. Houkin K, Nakayama N, Kuroda S, et al. Novel magnetic resonance angiography stage grading for moyamoya disease. Cerebrovasc Dis (Basel, Switzerland) 2005; 20:347-354. 11. Fazekas F, Kleinert R, Roob G, et al. Histopathologic analysis of foci of signal loss on gradient-echo T2*-weighted MR images in patients with spontaneous intracerebral hemorrhage: evidence of microangiopathy-related microbleeds. AJNR Am J Neuroradiol 1999;20:637-642. 12. Shoamanesh A, Kwok CS, Benavente O. Cerebral microbleeds: histopathological correlation of neuroimaging. Cerebrovasc Dis (Basel, Switzerland) 2011;32:528-534. 13. Okazaki S, Sakaguchi M, Hyun B, et al. Cerebral microbleeds predict impending intracranial hemorrhage in infective endocarditis. Cerebrovasc Dis (Basel, Switzerland) 2011;32:483-488. 14. Goulenok T, Klein I, Mazighi M, et al. Infective endocarditis with symptomatic cerebral complications: contribu-
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
32.
tion of cerebral magnetic resonance imaging. Cerebrovasc Dis (Basel, Switzerland) 2013;35:327-336. Brouwers HB, Greenberg SM. Hematoma expansion following acute intracerebral hemorrhage. Cerebrovasc Dis (Basel, Switzerland) 2013;35:195-201. Kuroda S. Cerebrovascular disease: new data on surgical therapy for pediatric moyamoya disease. Nat Rev Neurol 2010;6:242-243. Hamada J, Hashimoto N, Tsukahara T. Moyamoya disease with repeated intraventricular hemorrhage due to aneurysm rupture. Report of two cases. J Neurosurgery 1994;80:328-331. Irikura K, Miyasaka Y, Kurata A, et al. A source of haemorrhage in adult patients with moyamoya disease: the significance of tributaries from the choroidal artery. Acta Neurochir 1996;138:1282-1286. Kuroda S, Houkin K, Kamiyama H, et al. Effects of surgical revascularization on peripheral artery aneurysms in moyamoya disease: report of three cases. Neurosurgery 2001;49:463-467. discussion 467-468. Kobayashi E, Saeki N, Oishi H, et al. Long-term natural history of hemorrhagic moyamoya disease in 42 patients. J Neurosurg 2000;93:976-980. Nonaka H, Akima M, Hatori T, et al. Microvasculature of the human cerebral white matter: Arteries of the deep white matter. Neuropathology 2003;23:111-118. Moody DM, Bell MA, Challa VR. Features of the cerebral vascular pattern that predict vulnerability to perfusion or oxygenation deficiency: an anatomic study. AJNR Am J Neuroradiol 1990;11:431-439. Takahashi S, Goto K, Fukasawa H, et al. Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part ii: thalamic arterial group. Radiology 1985;155:119-130. Takahashi S, Goto K, Fukasawa H, et al. Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part i: striate arterial group. Radiology 1985;155:107-118. Kim JM, Lee SH, Roh JK. Changing ischaemic lesion patterns in adult moyamoya disease. J Neurol Neurosurgery Psychiatry 2009;80:36-40. Marinkovic S, Gibo H, Filipovic B, et al. Microanatomy of the subependymal arteries of the lateral ventricle. Surg Neurol 2005;63:451-458. discussion 458. Hosoda Y, Ikeda E, Hirose S. Histopathological studies on spontaneous occlusion of the Circle of Willis (cerebrovascular moyamoya disease). Clin Neurol Neurosurgery 1997;99(Suppl 2):S203-S208. De Reuck J. The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 1971;5:321-334. Nelson MD Jr, Gonzalez-Gomez I, Gilles FH. Dyke award. The search for human telencephalic ventriculofugal arteries. AJNR Am J Neuroradiol 1991;12:215-222. Van den Bergh R. The ventriculofugal arteries. AJNR Am J Neuroradiol 1992;13:413-415. Cordonnier C, Potter GM, Jackson CA, et al. Improving interrater agreement about brain microbleeds: development of the brain observer microbleed scale (bombs). Stroke 2009;40:94-99. Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol 2009;8:165-174.
Background: Adult moyamoya disease (MMD) is known to have high incidence of cerebral microbleeds (cMBs); however, the clinical significance still remains unclear. We investigated the frequency of cMBs in a large number of patients and analyzed the patterns of MB distribution in association with the location of the hematoma and moyamoya vessels. Methods: We studied 259 consecutive patients with MMD using prospectively collected database. One hundred ninety-one patients were eligible for the present study, and image analysis was performed retrospectively. The presence of cMBs and remains of hemorrhage were determined using gradient-echo T2*weighted sequence (1.5 T). The development of moyamoya vessels was assessed on source images of time-of-flight magnetic resonance angiography. The analysis consists of descriptive assessment of the spatial relationship between cMB, remains of hemorrhage, and moyamoya vessels. Statistical analysis was performed to calculate relative risk ratio in the presence of cMBs in relation to the remains of hemorrhage (macrohematoma), age of onset, and the presence of concomitant moyamoya vessels. Results: Thirty MBs were observed in 20 adult MMD patients (16.9%). MBs were located predominantly in the periventricular white matter (63.3%) followed by the basal ganglia/thalami (20%). Comparing the patients with cMBs from those without, hematoma was more frequently observed in patients with cMBs (odds ratio [OR] 4.29; 95% confidence interval [CI] 1.58-11.62; P 5 .0062). Patients with adult onset was more likely to demonstrate cMBs (14.4%) compared with the patients with pediatric onset (4.1%) (OR 3.93; 95% CI 1.11-13.91). Moyamoya vessels appeared in the lateral part of the trigon, and the periventricular white matter was significantly associated with the presence of cMBs (lateral part of the trigon; OR 3.29 [1.59-6.82], P 5 .0019, periventricle of the body of lateral ventricle; OR 2.40 [1.20-4.79], P 5.0214, respectively). cMBs accompanied concomitant arteries in 23 (76.7%) lesions. The subependymal–leptomeningeal artery anastomosis was the most common pattern (n 5 20, 66.7%). Conclusions: Spatial relationship was demonstrated between the moyamoya vessels and perivascular hemosiderin deposition particularly around the subependymal–leptomeningeal anastomosis, suggesting the mechanism for the development of cMBs in MMD. Present study further supports previous findings that cMBs potentially serve as a marker for the bleeding-prone microangiopathy in MMD. The significance of the present study lies in selecting optimal surgical candidate for preventing future hemorrhage by the presence of the cMBs, whereas current surgical indication relying on the degree
From the *Department of Neurosurgery, Hokkaido University Graduate School of Medicine, Sapporo; and †Department of Neurosurgery, Toyama University Graduate School of Medicine, Toyama, Japan. Received July 22, 2013; revision received October 10, 2013; accepted December 3, 2013. Funding: No grants or funding sources are pertinent to this article. Conflicts of interest: None.
Address correspondence to Ken Kazumata, MD, Department of Neurosurgery, Hokkaido University Graduate School of Medicine, North 15 West 7, Kita, Sapporo 060-8638, Japan. E-mail: kazumata@ med.hokudai.ac.jp. 1052-3057/$ - see front matter Ó 2013 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2013.12.007
Journal of Stroke and Cerebrovascular Diseases, Vol. -, No. - (---), 2013: pp 1-8
1
K. KAZUMATA ET AL.
2
of ischemia frequently fails to detect patients with future hemorrhage. Key Words: Moyamoya disease—microbleeds—intracranial hemorrhage—intraventricular hemorrhage—subependymal artery—medullary artery. Ó 2013 by National Stroke Association
Introduction MMD is a progressive steno-occlusive cerebrovascular disease that is characterized by the presence of net-like collateral vessels (moyamoya vessels) at the base of the brain.1 Recent reports demonstrated that the incidence of cerebral microbleeds (cMBs) was higher in patients with MMD compared with healthy subjects.2-6 In addition, the presence of cMBs predicts the future risk of hemorrhage.2,7 These results suggest that cMBs can serve as a marker of bleeding-prone microangiopathy in MMD. A certain pattern of distribution of MBs serves to indicate microangiopathy developed by the different mechanisms. cMBs in the basal ganglia and cortical– subcortical distribution are characteristic in hypertensive lipohyalinosis and amyloid angiopathy, respectively. Similarly, it is speculated that MMD may demonstrate specific pattern of the distribution of cMBs and follow the spatial distribution of hematoma and moyamoya vessels. The investigations of these spatial relationships may further elucidate the clinical significance and the mechanism for the development of the cMBs in MMD. In addition to the spatial distribution of cMBs, the exact number or locations of the bleeding have also not been fully evaluated previously. T2* gradient-echo or susceptibility weighted image not only demonstrate cMBs but also reveals remains of tiny intracerebral hemorrhage throughout life, which enables determining the cause of the hemorrhage years after the bleeding event. This is particularly useful to determine the location of the hemorrhage in primary intraventricular hemorrhage, a frequent manifestation of the intracranial bleeding in adult MMD. Furthermore, the source image of magnetic resonance angiography (MRA) demonstrates the enlarged basal perforators and choroidal arteries developed as a collateral circulation, which potentially reveal spatial association between the moyamoya vessels and the cMBs. Present study attempted to provide additional information by performing an analysis of the frequency and distribution patterns of MBs in association with the hematoma location and moyamoya vessels in a large consecutive series of patients with MMD.
Patients and Methods We examined 259 consecutive patients with MMD at the Department of Neurosurgery of the Hokkaido University Hospital from March 1980 to March 2012. The diagnosis was based on the guidelines established by the Research Committee on Moyamoya Disease (Spontaneous Occlu-
sion of the Circle of Willis) of the Ministry of Health and Welfare of Japan.8 These patients were referred to our hospital for the diagnosis and the treatment. Demographic information, including history of stroke, medical history, and medications for hypertension, diabetes mellitus, or hyperlipidemia, was collected. We also recorded the prevalence of antithrombotic therapy before the occurrence of recurrent stroke in each patient. Their information is tabulated in the database prospectively and used in this study to select eligible patients in this study. Since 2003, a standardized magnetic resonance imaging (MRI) protocol has been employed for all the time points that include fluid-attenuated inversion recovery (FLAIR), T2, and gradient recalled echo (GRE) sequences and MRA.5 Gradient-echo T2*-weighted sequence was available for assessment in 191 patients (73.7%; 56 men and 135 women; age, 2-86 years; mean age, 40.0 1 19.3 years). One hundred forty-one of the 191 patients underwent revascularization surgery (73.8%). In turn, 70 of these 141 patients underwent gradient-echo T2*-weighted sequence studies before revascularization surgery, and the remaining 71 underwent imaging after revascularization surgery. No patient had a known pathogenesis, such as head trauma, cerebral amyloid angiopathy, current anticoagulant therapy, arteriovenous malformation, cavernoma, or other systemic coagulation disorder.
Imaging Analysis MRI was performed every 1 year and/or at the time of recurrent stroke using a 1.5 T MR scanner (Magnetom Vision; Siemens Medical Solutions, Erlangen, Germany). The imaging protocol consisted of the axial spin-echo T1-weighted imaging (WI) (repetition time/echo time [TR/TE] 5 600/14 ms, number of excitation [NEX] 5 1, matrix size 5 512 3 512), axial fast spin-echo T2WI (TR/TE 5 4000/96 ms, effective echo train length 5 7, NEX 5 1, matrix size 5 512 3 512), and axial gradientrecalled echo T2*WI (TR/TE 5 800/26 ms, flip angle 5 20 , NEX 5 1, matrix size 5 256 3 256). All images were acquired with a 240-mm field of view, a 5.0mm section thickness, and a 1.5-mm intersection gap. Axial source images acquired by the 3-dimensional time-of-flight magnetic resonance angiography (3DTOF-MRA) (TR/TE 5 27/7.2 ms; flip angle 5 20 ; NEX 5 1, field of view 5 230 mm; recon matrix 5 512 3 512; number of slabs 5 4 [180 sections]; section thickness 5 .65 mm; acquisition time 5 9 min) were used for identifying moyamoya vessels.
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
Gradient-echo T2*-weighted sequence is also sensitive in detecting small remains of previous parenchymal hemorrhage in primary intraventricular hemorrhage. Definite cMBs were defined as rounded foci of hypointense signal on GRE sequences with a diameter of 2-10 mm and not well seen on T2WI.9 Microbleed mimics were carefully excluded. Vessels with linear or curvilinear lesions in the subarachnoid space, with a cortical or juxtacortical location, were excluded. We did not include lesions that were adjacent to the infarction or hemorrhage or the area of surgery. Areas of symmetric hypointensity in the globus pallidus or dentate nuclei on GRE MRI were considered to likely represent calcifications and were not included. Air– bone interfaces in the frontal and temporal lobes were excluded from the analysis. The topographical distribution was demonstrated on the basis of the Microbleed Anatomical Rating Scale.9 The diagnosis of moyamoya disease (MMD) was achieved using MRA in recent patients. Therefore, arterial involvement was described by the composite score of the rating based on TOF MRA published elsewhere instead of Suzuki’s grading.10 Moyamoya vessels with linear or curvilinear structures in the basal ganglia, thalamus, and periventricular white matter were assessed on the source image of TOF MRA. The presence was described by either ‘‘visible’’ or ‘‘not visible.’’ No attempt of quantitative assessment was carried out in terms of vascular density of the moyamoya vessels. For the current analysis, 2 investigators performed the imaging evaluations using most recent MRI acquisition in each individual. Presence of cMBs and hematoma and the location were investigated by 1 researcher (D.S.), whereas rating the scale of MRA and the location of moyamoya vessels were investigated by the other researcher (K.K.). We further consulted previous MRI acquisition to determine whether these are the de novo lesions. The process of data collection was blinded between the 2 investigators. After completing data collection, we assessed the radiological rating, and any disagreement between these authors was resolved in consensus meeting with a third author (I.M.).
Statistical Analysis Differences in dichotomous variables were analyzed using Pearson chi-squared analysis or Fisher exact test. Student t test or the ANOVA was used to analyze differences in the mean or median of continuous variables between groups.
Results Demographic Data The population used here included 118 patients with adult MMD (age at onset, .18 years) and 73 patients with pediatric MMD (age at onset, #18 years). The cohort included presence of hematoma (n 5 31), presence of
3
cMBs alone (n 5 12), and absence of hematoma and cMBs (n 5 148). The average follow-up period was 104.0 6 92 months. Thirty cMBs were detected in 20 patients (age range 2365 years) on gradient-echo T2*-weighted sequence, corresponding to 16.9% in adult population; 15 cMBs were found in 8 patients with hematoma, whereas the remaining 15 cMBs were found in 12 patients without hematoma (ischemic symptoms [n 5 7], asymptomatic [n 5 4], and headache [n 5 1]; P , .01). Comparing the patients with cMBs from those without cMBs, hematoma was significantly more frequent in the patients with cMBs (odds ratio [OR] 4.29; 95% confidence interval [CI] 1.58-11.62; P 5 .0062). Patients with adult onset was more likely to demonstrate cMBs (14.4%) compared with the patients with pediatric onset (4.1%) (OR 3.93; 95% CI 1.11-13.91). The clinical parameters of patients in the groups with cMBs and no cMBs are listed in Table 1.
Topographic Patterns of cMBs and Hematoma In total, 41 macrohemorrhage (hematoma and tiny remains of hemorrhage at the paraventricle) lesions in 31 patients were identified (Fig 1, Table 2). Primary ventricular hemorrhage was documented in 8 patients before using gradient-echo T2*-weighted sequence. Small remains of hemosiderin on gradient-echo T2*-weighted sequence identified the origin of intraventricular hemorrhage (IVH) as being subependyum of the trigon (n 5 1), the lateral ventricle (n 5 2), and the temporal horn (n 5 1). cMBs were subependymal to the lateral ventricle in the lateral part of the trigon or the posterior two thirds of the periventricle (deep and periventricular white matter) in 19 lesions (63.3 %). Lobar microbleeds were observed in 3 lesions. Locations classified according to the Microbleed Anatomical Rating Scale are demonstrated in Table 2 and Figure 1. A trend for susceptibility to bleeding was observed in the basal ganglia and corpus callosum (P 5 .0184). The deep and periventricular white matter was susceptible to the development of cMBs (P 5 .07).
Moyamoya Vessels on Source MRA Images and Their Relationship with cMBs and Hemorrhage On MR, the linear structures were oriented in the known direction of perforating arteries (Figs 2-4). The source images of MRA were available in total of 298 hemispheres (cMBs [n 5 38], no cMBs [n 5 260]). In adult, moyamoya vessels were visible in the basal ganglia (n 5 78, 26.2%), thalamus (n 5 47, 15.8%), lateral part of the trigon (n 5 58, 19.5%), and periventricular white matter (n 5 89, 29.9%). Moyamoya vessels appeared in the lateral part of the trigon, and the periventricular white matter was significantly associated with the presence of cMBs (lateral part of the trigon; OR 3.29 [1.59-6.82], P 5 .0019, periventricle of the body of
K. KAZUMATA ET AL.
4
Table 1. Demographic and clinical variables for patients with cMBs and without cMBs Variables
cMBs (n 5 20)
No cMBs (n 5 171)
Age, y (mean 6 SD) Male, n (%) Symptom, n (%) Headache TIA Infarction Surgery, n (%) Onset age . 18 y, n (%) Medication Hypertension Hyperlipidemia Antiplatelet Hematoma, n (%) MRA composite score*
39.0 6 19.7 4 (20.0)
48.0 6 12.6 47 (27.5)
0 7 (35.0) 1 (5.0) 10 (50.0) 17 (85.0)
7 (4.1) 77 (45.0) 45 (26.3) 116 (67.8) 101 (59.1)
6 (30.0) 1 (5.0) 3 (15.0) 8 (40.0) 10.5 6 3.9
32 (18.7) 11 (6.4) 21 (12.3) 23 (13.5) 11.3 6 4.6
OR
P value ,.001 NS
3.93 (95% CI 1.11-13.91)
NS NS NS NS .0028
4.29 (95% CI 1.58-11.62)
NS NS NS .0062
Abbreviations: CI, confidence interval; cMBs, cerebral microbleeds; MRA, magnetic resonance angiography; NS, nonsignificant; OR, odds ratio; TIA, transient ischemic attack. *Arterial involvement rated by using MRA10
lateral ventricle; OR 2.40 [1.20-4.79], P 5 .0214, respectively). The typical patterns of anastomosis were identified in relationship with the moyamoya vessels as follows:
a) Distal part of the lateral posterior choroidal artery terminates in subependymal artery (SEA). SEA penetrated the lateral part of the trigon and the mid-posterior part of the lateral wall of the lateral
Figure 1. The patterns of distribution of microbleeds (open circles) and hemorrhage (closed circles) in moyamoya disease were depicted on the Microbleed Anatomical Rating Scale template.9
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
Table 2. The number of cMBs and the hemorrhage classified based on the MARS
Basal ganglia Thalamus Internal capsule External capsule Corpus callosum Deep and periventricular white matter Frontal Parietal Temporal Occipital Insula Brain stem Cerebellum
MBs
Hemorrhage
2 2 2 0 0 19
13* 3 4 0 7* 12y
2 1 0 0 2 0 0 30
0 0 1 0 0 1 0 41
Abbreviations: cMBs, cerebral microbleeds; MARS, Microbleeds Anatomical Rating Scale. Prevalence of cMBs and hematoma were compared across the regions. Hematoma is frequently observed in the basal ganglia and the corpus callosum, whereas cerebral microbleeds is relatively rare in these locations. cMBs are frequently observed in the deep and periventricular white matter (P 5.63), whereas hematoma in deep and periventricular white matter is less frequent manifestation (P 5.29). *P , .05, statistically significant. yP , .01, statistically significant, by the Fisher exact test.
ventricle. The linear signal took a radiating course toward the deep sulci (medullary artery), where leptomeningeal arteries in the temporoparietal region were identified (Fig 2).
Figure 2. Typical periventricular microbleeds and subependymal–leptomeningeal anastomosis. (A) Gradient-echo T2*-weighted imaging at the level of the lateral ventricles. (B and C) The source image of time-of-flight magnetic resonance angiography shows a linear high signal in the subependymal zone, deep white matter (short arrow), and deep sulci, which is consistent with a collateral pathway that developed because of anastomosis between subependymal artery and the leptomeningeal artery. (D and E) The coronal images in different patient demonstrate the course of the enlarged medullary arteries developed in bilateral parietal regions.
5
b) Thalamoperforating or medial posterior choroidal artery penetrated the lateral wall of the third ventricle. These anastomotic arteries traverse the basal ganglia and thalami, generating an anastomosis with either the long insular artery or the medullary artery (Fig 3). c) Anastomosis between the lenticulostriate artery and the medullary artery (Fig 3). In 23 (76.7%) cMBs, moyamoya vessels are identified lying next to them. Of these, anastomosis between SEA and the leptomeningeal artery was identified in 20 cases (66.7%). In 4 of the basal ganglia and thalamic cMBs (striatal; n 5 3, thalamus; n 5 20), a concomitant artery was identified in 4 cases (SEA–medullary artery, n 5 3; lenticulostriate artery, n 5 1). Source images of TOF MRA revealed concomitant arteries lying next to the remains of hematoma in 17 of 29 patients (58.6%) and concomitant arteries in 11 lesions of 9 patients with striatal hemorrhage. SEA–medullary anastomosis was identified in 5 of 8 patients with primary intraventricular hemorrhage. Source images of TOF MRA acquired before the hemorrhagic event were available for 8 of 10 cases of hemorrhage and 5 of 6 de novo cMBs. Retrospective image assessment identified moyamoya vessels at the site of 6 of 8 subsequent hemorrhage and 5 of 5 de novo cMBs.
Discussion In this study, we investigated the prevalence and distribution pattern of the cMBs among the largest number of patients with MMD investigated to date. The 17% rate
6
Figure 3. Patterns of moyamoya vessels. The source image of magnetic resonance angiography demonstrates the spatial relationship of the basal ganglia, lateral ventricle, and moyamoya vessels in coronal views. The lenticulostriate artery develops and forms an anastomosis between the medullary artery at the corner of the lateral ventricle on the right hemisphere (arrow; upper image). Bilateral thalamoperforating arteries give rise to the collateral network with long insular and medullary arteries on the left hemisphere (small arrow; middle 2 images). Choroidal arteries develop and form anastomosis with the medullary arteries at the lateral corner of the trigon (small arrow; bottom image).
of cMBs in MMD was lower than our preceded prospective cohort study in the nontreated MMD. We extended our survey to obtain greater population even in patients with a history of the revascularization surgery. The patients with cMBs were observed predominantly in adult-onset MMD. Posterior two thirds of the periventricular white matter of the lateral ventricle was the most common location of the cMBs. Presence of the periventricular moyamoya vessels and the hematoma is associated with the greater frequency of the cMBs. In addition, the location of the periventricular cMBs were considered to correspond with the site of anastomosis developed between the SEA and the medullary artery. cMBs are iron or hemosiderin deposits in the macrophages located in the perivascular space adjacent to small arteries, arterioles, and capillaries.11,12 It is considered as a marker of microangiopathy and speculated as a smallest form of the hemorrhage in some investigators.13-15 The average age of patients was 44 years old with cMBs,
K. KAZUMATA ET AL.
10 years younger than that of patients with the history of hemorrhage (54 years old). cMBs were more frequent in patients with hematoma than in patients without. Adult patients who underwent revascularization in childhood were less likely to demonstrate cMBs. These results suggest that cMBs may develop in adult MMD with certain disease duration and are associated with the later hemorrhage. The revascularization surgery might have prevented from developing these bleedingprone microangiopathy.16 The study of the distribution may further elucidate the clinical significance of cMBs. The periventricular predominant cMBs is considered characteristic distribution of MMD because deep periventricular cMBs are rare in stroke patients with other etiologies.9 In MMD, intraventricular hemorrhage is a frequent manifestation of the intracranial bleeding. Previously, the microaneurysm arising at the choroidal arteries have been considered responsible for such bleeding.17-19 Gradient-echo T2*weighted sequence revealed tiny remains of hematoma in periventricle in patients with a history of the primary hemorrhage. This suggests that the cause of IVH has 2 distinct mechanisms: one is the rupture of intraventricular aneurysms and the subependymal hemorrhage ruptured into the ventricle in the other. The latter is considered to take more benign prognosis than the former because the performance state in patients with IVH was not as poor as previously reported.20 The hemorrhage in the basal ganglia is observed in approximately half of the intracranial hemorrhage, whereas the cMBs in the basal ganglia and thalami were observed in only 20%. This discrepancy indicates that a certain population of hemorrhage in the basal ganglia develops spontaneously. In addition, it is speculated that the culprit vessels for the hemorrhage and cMBs may not be entirely the same in the basal ganglia. We further investigated the spatial relationship between the moyamoya vessels and the cMBs. The medullary artery irrigates white matter above the level of the lateral ventricle via the leptomeningeal artery.21 In particular, cortical arteries located around the watershed zone between the anterior cerebral and middle cerebral artery give rise to the medullary arteries.22 Lenticulostriate and long insular arteries irrigate white matter below the level of lateral ventricle.23,24 A collateral network via the deep perforating arteries or choroidal arteries develops when the cortical arteries were involved with disease progression. Lenticulostriate, thalamoperforating, and medial posterior choroidal arteries participate in the collateral networks with medullary arteries in the frontal region. Lateral posterior choroidal artery participates in the collateral network in the temporoparietal region. These anastomotic networks develop most exclusively at the lateral corner to the lateral ventricle. In patients with advanced age, temporoparietal region is susceptible for the ischemia.25 Therefore, a collateral network between
RELATIONSHIP BETWEEN cMBs, MOYAMOYA VESSELS, AND HEMATOMA IN MMD
7
Figure 4. Illustrations of the hemorrhagic pattern attributed to bleeding from the subependymal artery (SEA)–leptomeningeal anastomosis. (A) Subependymal hemorrhage can be regarded as an enlarged form of periventricular cerebral microbleeds (cMBs). (B) A case of intraventricular hemorrhage. Remains of hemorrhage can be traced along the medullary artery. (C) cMBs in the posteroventral putamen developed along the course of SEA–medullary anastomosis. (D) Subcortical hemorrhage in the temporal lobe that developed because of bleeding from the SEA–medullary anastomosis.
the posterior choroidal and the medullary artery is considered developing predominantly in adult patients with advanced age. This may explain that cMBs develop most frequently at the site of anastomosis taking place between the choroidal arteries, deep perforators, and the medullary arteries. A 5- to 30-fold increase in diameter is estimated when SEA or the medullary artery are visible on MRI as the size of the choroidal arteries is 190 mm on average (range 40-490 mm) and the size of arterioles is 20-30 mm.26 Pathological studies of collateral circulation have demonstrated the presence of hemodynamic stress in MMD, revealing the breakdown of the elastic lamina and the thinning of the media.27 The pattern of collateralization via the SEA can also be described as a ventriculofugal and ventriculopetal anastomosis.28-30 MMD is the first of its kind to demonstrate such anastomosis clearly in living human imaging. The pattern of hematoma observed in MMD suggests that SEA–leptomeningeal anastomosis potentially induce a wide array of hemorrhagic lesions located in the subependymal zone, ventricle, posterior ventral putamen, and temporal subcortical region (Fig 4). The limitations of this study include its retrospective nature and the modality used, which is less sensitive compared with the currently available high-field MR scanner.4 In contrast with previous prospective studies conducted using a high-field MR scanner, this retrospective study revealed a smaller incidence of cMBs in
MMD.2,4,5,7 It should be noted that the signal intensity observed in blood vessels, as imaged by threedimensional TOF MRA, is dependent on the rate of blood flow. The frequency accompanying a concomitant artery adjacent to cMBs is, thus, potentially inaccurate and may be increased in studies using a high-field modality. It is necessary to differentiate gradient-echo T2*weighted hypointense lesions from cMB mimics, such as small vessels, calcification of small arteries, distended dissecting arteries, thrombosed arterioles, and capillaries. In particular, subependymal cMBs can be seen as an intraventricular lesion that mimics hemosiderin deposition that occurs in the choroid plexus after intraventricular hemorrhage. Therefore, some of these subependymal cMBs may be regarded as probable cMBs.9,31 However, considering the diameter of the cMBs, which was larger than the size of the accompanying vessels, and the absence of flow void on T2WI, gradient-echo T2*weighted hypointense lesions are consistent with the findings of cMBs.9,32
Conclusions The prevalence and distribution of cMBs were studied in a large number of patients with MMD. The distribution of cMBs in MMD showed a strong predilection for the subependymal zone of the posterior two thirds of the
K. KAZUMATA ET AL.
8
lateral ventricle. Collateral circulation between SEA and the medullary artery was demonstrated in close proximity to cMBs, suggesting that anastomosis between the ventriculopetal and ventriculofugal arteries is associated with the characteristic distribution of cMBs in MMD.
References 1. Suzuki J, Takaku A. Cerebrovascular ‘‘moyamoya’’ disease. Disease showing abnormal net-like vessels in base of brain. Arch Neurol 1969;20:288-299. 2. Kikuta K, Takagi Y, Nozaki K, et al. Asymptomatic microbleeds in moyamoya disease: T2*-weighted gradientecho magnetic resonance imaging study. J Neurosurgery 2005;102:470-475. 3. Kikuta K, Takagi Y, Nozaki K, et al. Histological analysis of microbleed after surgical resection in a patient with moyamoya disease. Neurol Med Chir 2007;47:564-567. 4. Kikuta K, Takagi Y, Nozaki K, et al. The presence of multiple microbleeds as a predictor of subsequent cerebral hemorrhage in patients with moyamoya disease. Neurosurgery 2008;62:104-111. discussion 111-102. 5. Ishikawa T, Kuroda S, Nakayama N, et al. Prevalence of asymptomatic microbleeds in patients with moyamoya disease. Neurol Med Chir 2005;45:495-500. discussion 500. 6. Kuroda S, Kashiwazaki D, Ishikawa T, et al. Incidence, locations, and longitudinal course of silent microbleeds in moyamoya disease: a prospective t2*-weighted MRI study. Stroke 2013;44:516-518. 7. Mori N, Miki Y, Kikuta K, et al. Microbleeds in moyamoya disease: susceptibility-weighted imaging versus T2*-weighted imaging at 3 Tesla. Invest Radiol 2008; 43:574-579. 8. Fukui M. Guidelines for the diagnosis and treatment of spontaneous occlusion of the Circle of Willis (’moyamoya’ disease). Research committee on spontaneous occlusion of the Circle of Willis (moyamoya disease) of the ministry of health and welfare, Japan. Clin Neurol Neurosurgery 1997;99(Suppl 2):S238-S240. 9. Gregoire SM, Chaudhary UJ, Brown MM, et al. The Microbleed Anatomical Rating Scale (MARS): reliability of a tool to map brain microbleeds. Neurology 2009; 73:1759-1766. 10. Houkin K, Nakayama N, Kuroda S, et al. Novel magnetic resonance angiography stage grading for moyamoya disease. Cerebrovasc Dis (Basel, Switzerland) 2005; 20:347-354. 11. Fazekas F, Kleinert R, Roob G, et al. Histopathologic analysis of foci of signal loss on gradient-echo T2*-weighted MR images in patients with spontaneous intracerebral hemorrhage: evidence of microangiopathy-related microbleeds. AJNR Am J Neuroradiol 1999;20:637-642. 12. Shoamanesh A, Kwok CS, Benavente O. Cerebral microbleeds: histopathological correlation of neuroimaging. Cerebrovasc Dis (Basel, Switzerland) 2011;32:528-534. 13. Okazaki S, Sakaguchi M, Hyun B, et al. Cerebral microbleeds predict impending intracranial hemorrhage in infective endocarditis. Cerebrovasc Dis (Basel, Switzerland) 2011;32:483-488. 14. Goulenok T, Klein I, Mazighi M, et al. Infective endocarditis with symptomatic cerebral complications: contribu-
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30. 31.
32.
tion of cerebral magnetic resonance imaging. Cerebrovasc Dis (Basel, Switzerland) 2013;35:327-336. Brouwers HB, Greenberg SM. Hematoma expansion following acute intracerebral hemorrhage. Cerebrovasc Dis (Basel, Switzerland) 2013;35:195-201. Kuroda S. Cerebrovascular disease: new data on surgical therapy for pediatric moyamoya disease. Nat Rev Neurol 2010;6:242-243. Hamada J, Hashimoto N, Tsukahara T. Moyamoya disease with repeated intraventricular hemorrhage due to aneurysm rupture. Report of two cases. J Neurosurgery 1994;80:328-331. Irikura K, Miyasaka Y, Kurata A, et al. A source of haemorrhage in adult patients with moyamoya disease: the significance of tributaries from the choroidal artery. Acta Neurochir 1996;138:1282-1286. Kuroda S, Houkin K, Kamiyama H, et al. Effects of surgical revascularization on peripheral artery aneurysms in moyamoya disease: report of three cases. Neurosurgery 2001;49:463-467. discussion 467-468. Kobayashi E, Saeki N, Oishi H, et al. Long-term natural history of hemorrhagic moyamoya disease in 42 patients. J Neurosurg 2000;93:976-980. Nonaka H, Akima M, Hatori T, et al. Microvasculature of the human cerebral white matter: Arteries of the deep white matter. Neuropathology 2003;23:111-118. Moody DM, Bell MA, Challa VR. Features of the cerebral vascular pattern that predict vulnerability to perfusion or oxygenation deficiency: an anatomic study. AJNR Am J Neuroradiol 1990;11:431-439. Takahashi S, Goto K, Fukasawa H, et al. Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part ii: thalamic arterial group. Radiology 1985;155:119-130. Takahashi S, Goto K, Fukasawa H, et al. Computed tomography of cerebral infarction along the distribution of the basal perforating arteries. Part i: striate arterial group. Radiology 1985;155:107-118. Kim JM, Lee SH, Roh JK. Changing ischaemic lesion patterns in adult moyamoya disease. J Neurol Neurosurgery Psychiatry 2009;80:36-40. Marinkovic S, Gibo H, Filipovic B, et al. Microanatomy of the subependymal arteries of the lateral ventricle. Surg Neurol 2005;63:451-458. discussion 458. Hosoda Y, Ikeda E, Hirose S. Histopathological studies on spontaneous occlusion of the Circle of Willis (cerebrovascular moyamoya disease). Clin Neurol Neurosurgery 1997;99(Suppl 2):S203-S208. De Reuck J. The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 1971;5:321-334. Nelson MD Jr, Gonzalez-Gomez I, Gilles FH. Dyke award. The search for human telencephalic ventriculofugal arteries. AJNR Am J Neuroradiol 1991;12:215-222. Van den Bergh R. The ventriculofugal arteries. AJNR Am J Neuroradiol 1992;13:413-415. Cordonnier C, Potter GM, Jackson CA, et al. Improving interrater agreement about brain microbleeds: development of the brain observer microbleed scale (bombs). Stroke 2009;40:94-99. Greenberg SM, Vernooij MW, Cordonnier C, et al. Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol 2009;8:165-174.
